Universal Access To All Knowledge
Home Donate | Store | Blog | FAQ | Jobs | Volunteer Positions | Contact | Bios | Forums | Projects | Terms, Privacy, & Copyright
Search: Advanced Search
Anonymous User (login or join us)
Upload

Reply to this post | See parent post | Go Back
View Post [edit]

Poster: Scribble Date: Apr 9, 2012 12:37am
Forum: prelinger Subject: Re: New Films

Yes, I have to confirm these observations...I've never had too much trouble removing pulldown from the old MPEG2 encoded films...but these new mp4s, while I must say look terrific resolution wise, have a pulldown cadence that does not conform to any normal pattern, there is definitely frame blending going on and even without trying to remove pulldown, plays back quite jittery at 29.97. Just as Frank notes, it's hard to complain regarding a free resource, I am eternally grateful, but it's a real shame that clean, 24fps, or even 30fps slowed down to 24fps, are not happening with this new round of footage...

Reply to this post
Reply [edit]

Poster: Frank Panucci Date: Apr 9, 2012 10:45am
Forum: prelinger Subject: Re: New Films

I keep a number of Prelinger-supplied films on my hard drive permanently because I use footage from them so often. So, I have conveniently at hand for side-by-side comparison the WESTINGHOUSE TRAVELERS' CHOICE 66 MPEG2 encode I downloaded eight years ago, and the MP4 encode uploaded recently.

MPEG2 has two channels of sound.
MP4 has only a left channel and a blank right.
It seems to be mono sound anyway, but the empty channel is confusing.

MPEG2 audio is sharp and distinct.
MP4 audio is slightly muffled and distorted. Its waveform reveals some degree of additional compression compared to the MPEG2 audio.

MPEG2 is half NTSC horizontal resolution with faded color.
MP4 has better color, and is sharper, at full 720 res.
When things aren't moving, the image is very good.

MPEG2 is a true 60i video file with proper pulldown, which can be loaded in a video program and broken out into a 24fps progressive image sequence.
MP4 is a 29.97 progressive file, with blended fields, which very much appear to have been derived from a reversed-field transfer. Almost any motion is severely affected. Pulldown can absolutely not be removed.

I prepared a side-by-side screen grab of the same frame from both encodes. Maybe it will help diagnose problems on the encoding end for future Prelinger offerings:
http://img850.imageshack.us/img850/4939/prelingerartifact2012b.jpg


This post was modified by Frank Panucci on 2012-04-09 17:43:17

This post was modified by Frank Panucci on 2012-04-09 17:45:41

Reply to this post
Reply [edit]

Poster: Frank Panucci Date: Apr 9, 2012 8:20am
Forum: prelinger Subject: Re: New Films

http://archive.org/details/0657_Carving_Magic
CARVING MAGIC

Same issues as described above, compared to the original upload in 2003 or so.

Reply to this post
Reply [edit]

Poster: Frank Panucci Date: Apr 9, 2012 8:24am
Forum: prelinger Subject: Re: New Films

http://archive.org/details/0664_HM_Medicus_collection_New_York_Worlds_Fair_1939-40_Reel_2
Medicus collection: New York World's Fair, 1939-40 (Reel 2)

Same technical glitches as above, compared to the original encode of the same film also available here.

I'm not trying to be a pain, really. I'm sure you would prefer to post films as watchable and technically useful as those for which the Prelinger-branded section of the archive has been known for years. The number of improperly-encoded videos is growing larger at an upsetting rate of speed.




Reply to this post
Reply [edit]

Poster: Frank Panucci Date: Apr 9, 2012 9:30am
Forum: prelinger Subject: Re: New Films

I checked over my recent Prelinger backups. As a matter of routine I download every film at archive.org that looks remotely interesting, in case a useful gem is taken down for some reason, or if the archive itself becomes unavailable. I often don't examine the files for weeks or months. I also checked over another hundred or so entries I hadn't downloaded yet.

It appears, unfortunately, that the Prelinger-section films afflicted with the technical errors detailed above number over four hundred, and include everything posted since December 2011.

Reply to this post
Reply [edit]

Poster: Administrator, Curator, or StaffRick Prelinger Date: Apr 9, 2012 3:40pm
Forum: prelinger Subject: Re: New Films

I'm just now seeing the new posts in this thread. We are looking into this and will try to figure out where this is happening.

Reply to this post
Reply [edit]

Poster: Scribble Date: Apr 10, 2012 6:54am
Forum: prelinger Subject: Re: New Films

Rick, I have to again agree with Frank, the visual quality of the new mp4s is undeniably better, a vast improvement over the mpegs, but yes, unfortunately I am getting exactly the same pull-down issues that Frank has so perfectly described. Luckily, I have many of the mpeg2 versions of films I make use of which, unfortunately despite their inferior quality, I would be more likely to use because of their correct pulldown which is very easy to remove.

I really appreciate that this resource has been available for so many years, and yes, I hate being a pain about it, but it would be so incredibly valuable for low/no budget filmmakers to have these files with either proper pulldown...or, even better, no pulldown at all. It would make the files smaller and many productions are being edited at 24fps these days...with pulldown added at the end of the process.

Reply to this post
Reply [edit]

Poster: Administrator, Curator, or StaffRick Prelinger Date: Apr 23, 2012 7:53am
Forum: prelinger Subject: Re: New Films

Just to be clear, what we're doing is digitizing tapes of telecine transfers made between 1984 and 2011. All are NTSC SD transferred at 29.97 fps, but beyond that they could be anything; they were transferred by various facilities at various times on various machines. While we can control the parameters by which we derive mp4s to upload, the transfers are what they are.

Reply to this post
Reply [edit]

Poster: philips272 Date: Jun 2, 2012 3:30pm
Forum: prelinger Subject: Re: New Films

The quality is much better but this clip appears to be reel 3 according to the older mpeg file